Dataset: Dimensionality of infection networks among viruses infecting microbial eukaryotes and bacteria from Edwards and Hayward (2024)

The supplemental .zip packages containing R-code that performed the analysis and input files containing the compiled sources are available from the publication (Edwards, K. (2024), https://doi.org/10.6084/m9.figshare.24991041).

"R Code for analyzing network dimensionality" supplementary_file_S1.zip contains three scripts that analyze the dimensionality of infection networks.   

(1) The script ‘for_publication_nullmodel.R’ takes an infection matrix and uses a null model to test whether the infection matrix can be sorted in one dimension better than expected by chance. As noted in the script, the script assumes that the matrix is a csv file formatted such that rows correspond to viruses, columns correspond to hosts, the first column contains virus strain labels, and the first row contains host strain labels. 1 = infection, 0 = no infection. The script was written in R 4.3.2 and also uses a package called future.apply for parallelization.

(2) The script ‘for_publication_find_minimum_dimensionality.R’ takes and infection matrix and finds the smallest number of dimensions in which an infection matrix can be sorted with zero errors. The data input is the same as the null model script. 

(3) The script ‘for_publication_dimensionality_plot_functions.R’ contains plotting functions that are referenced by the second script. 

Other details of how the code works are included as annotations within the scripts themselves. 

"Phytoplankton-virus infection networks" supplementary_file_S2.zip contains 11 phytoplankton-virus networks stored as .csv files, where the rows are virus strains and the columns are host strains. Within the matrix a value of 1 indicates that the virus strain can infect the phytoplankton strain, while 0 indicates that it cannot. The csv files are named with the last name of the first author of the publication, followed by the date of publication.